the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 30 Sep 2024
Intermodel comparison of the atmospheric composition changes due to emissions from a future supersonic aircraft fleet
Abstract. Commercial supersonic aircraft may return in the near future, offering reduced travel times while flying higher in the atmosphere than present-day aircraft. Their emissions can change the composition of the atmosphere, particularly in the concentration and spatial distribution of ozone, aerosols, and greenhouse gases, posing risks to both the climate and public health. We present a comprehensive multi-model assessment of the impact of a supersonic fleet on a 2050 atmosphere using four state-of-the-art atmospheric models (EMAC, GEOS-Chem, LMDZ-INCA, MOZART-3). We show that the adoption of a fleet with a NOx emissions index of 4.6 g(NO2)/kg(fuel) leads to a model-mean stratospheric H2O burden of +61.34 Tg for 46.2 Tg of annual H2O emissions, and an ozone column loss of -0.11 %. With a NOx emissions index of 13.8 g(NO2)/kg(fuel) the average ozone column loss increases to -0.31 %. A lower cruise altitude and speed reduces the mean H2O burden to +9.34 Tg and instead leads to an ozone column increase of +0.02 %. Compared to the most recent multi-model assessment (2007), we find better agreement between the models, especially for the ozone response. Disagreement in H2O perturbation lifetimes remains, potentially driven by differences in vertical model resolutions. Our results reaffirm that emissions from a supersonic aircraft fleet will lead to global changes in atmospheric composition, which can be reduced by adopting lower cruise altitudes and lowering NOx emissions.
- Preprint
(4450 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2024-2866', Anonymous Referee #1, 29 Oct 2024
Van ‘t Hoff et al. presented a multi-model analysis of the impact of a future supersonic aircraft fleet on upper troposphere and stratospheric composition. Overall, it is a good paper and the results provide valuable information for future consideration of deployment of commercial supersonic aircraft. However, I have some major comments, as well as some minor comments, that need to be addressed before the paper is accepted for publication. Here are the specifics.
Major comments:
- Section 2.2. The vertical resolution of these models in the upper troposphere and lower stratosphere, could be a critical configuration setup in understanding the aviation impact differences, e.g. Table 3 and Figure 2. Could you elaborate the vertical resolution of each model? For example, how many layers in the critical region (200 hPa – 50 hPa), average layer thickness?
- One of the most interesting findings from this work is the distinctive responses in NOx, O3, and H2O in EMAC which is an online vs the other three models with nudged meteorology. As the authors mentioned in the text, this is likely due to the meteorological/dynamical feedback induced by the composition changes in H2O and O3. One major weakness of this work is that while the authors have presented many figures showing the chemical responses both in the main text and in supplement, I couldn’t find any showing the dynamical changes, e.g. temperature, U/V/omega, other transport terms, in EMAC due to supersonic aircraft emissions w.r.t. the control run. It would be useful to include these results as they will help to elucidate the changes in H2O, NOx, NOy, and ozone in different parts of the atmosphere.
- Section 3.6. This is a good discussion section, but I would suggest, instead of “modeling consideration”, make it a more comprehensive discussion to include the following aspects: i) what are the key findings from this paper? Ii) how these findings agree/differ from previous work (Grewe et al., 2007; Zhang et al., 2023; Eastham et al 2022; Matters et al., 2022), qualitatively and quantitatively? iii) what are the implications in differences in offline CTMs vs. online meteorology (as in EMAC)? For example, does the results in this work suggest that models with offline meteorology are possibly not the optimal way to assess supersonic aircraft emissions impact as the offline setup is inadequate in capturing full responses. Or alternatively, you can state what are the key impacts that can be captured by offline CTM and what are the key impacts that occur in atmosphere but missing in offline CTM representation.
- Other than the technical documentation of composition changes, what is the bigger implication of this study? Can the authors add a few sentences in the conclusion section about the implication? As presented throughout the paper, the adoption of A1 scenario supersonic aircraft fleet leads to very small changes in H2O (1-2%), NOx (1-3%), O3 (a few tenths percent, <1DU). The A3 lower cruise scenario impact is even smaller. Does this suggest it is reasonable to consider supersonic aircraft fleet in the future?
Minor comments:
- Title and elsewhere in the text. It is more common to use “multimodel intercomparison” instead of “intermodal comparison”. Consider revise.
- L17 & L96. Other than EMAC, I would not agree to use the word “state-of-the-art” to describe the remaining three models, especially in the context of this paper here. The CTM capability does not allow to examine the meteorological feedback of the composition change induced by the aircraft emissions. You can use a softened word such as “comprehensive atmospheric models”.
- L19-20. Here and in most of the places through the text, you are presenting the ozone changes in % unit. It would be much more helpful to me if you express these changes in dobson unit (DU), or at least include the DU changes in parenthesis, which is the common unit used in literature.
- L28. While it isn’t wrong to use “intercontinental transport”, to not be confused with intercontinental transport of air mass or pollutants (which is another major topic in the composition community), I would suggest you use “intercontinental transportation” instead.
- L141. Should be Burkholder et al.
- L196-199. Please clarify what is the physical water vapour tracer vs. full water vapour tracer, how are they treated differently in the model.
- L208. Missing parenthesis after 2011.
- L215. What do you mean by “fixing it the troposphere”? Relaxed to climatology, or nudge with ERA-Interim data?
- L237. Change “Firstly” to First. Also if you are listing First, shouldn’t you also use Second and Third, etc. ? I don’t see any in the following sentences.
- L238-239. This sentence seemed to be out of place. Probably better suited somewhere in Section 2.2.
- L245 and thereafter – the use of term “hemispheric fraction”. As you defined in the text, this is the ratio of the perturbation between the two hemispheres, not the fraction. By mathematical definition, a fraction is a numerical quantity that represent the portion/part of the whole thing and varies between 0.0 and 1.0. Please use a more accurate term to describe this.
- L247. Can you actually learn about interhemispheric transport, or is it just the NH --> SH transport since aviation emissions are pre-dominantly emitted in the NH, particularly in the NH mid-latitudes?
- Section 3.1. I find this section is a little hard to follow as the authors are frequently hoping back and forth from model mean responses to inter-model differences, and vs. previous results from Grewe et al. (2007). May be consider re-arrange the discussion in the following order: model mean impact, including H2O, NOx, Ozone in A1 scenario, A2 scenario and A3 scenario--> how do the A1 results compare with Grewe et al. (2007). You may want to mention model spread here, but I find it distracting if you discuss the details of why one model is different from the others. You can move the discussions into the later sections.
- L 257. I am not convinced that the authors presented conclusive evidence to suggest that a smaller NH/SH ratio is due to reduced hemispheric exchange. Did you check the changes in H2O lifetime? Short H2O lifetime could lead more faster removal of H2O (from aviation emissions) when it is en route from the NH to SH? There are other possible explanations as well.
- L277. What is the reason behind a net-decrease in H2O in the SH? Is it due to temperature perturbation? Is it due to transport responses from the composition changes? You may consider analyze the model temperature fields, vertical and horizontal transport terms to understand the causes.
- Figure 2. Are the dash-dotted lines model tropopause from each model? Please clarify in the figure caption.
- L312. I am not sure you should refer to the inter-model differences as “discrepancies”, large variances may be? By discrepancy, you are implying something is unexpected or some models may be inadequately wrong.
- L314. Change “afterwards” to (section xx).
- L315-327. See my comment above on model vertical resolution discussion.
- Figure 3. I would suggest you use the same value range for x-axis on all three panels. This way one can clearly identify the magnitude of differences in H2O changes.
- L322. Results from model are not Observed! Models are not observations. You may use alternative words such as shown, found, identified, etc.
- L329. “This is reflected ...” What is reflected? Please be specific.
- L354. Change “fit” to “agree with”
- L385. Delete “be”. Also slow down was used twice in this sentence.
- L392. As I have stated in previous comments, it would be useful to look at the meteorological changes in EMAC to confirm whether the unique responses in EMAC is due to online meteorology.
- L400. Change to “These effects combined”
- Figure 8 caption: change “changes in percentage of the ozone columns” --> “changes in ozone columns (in percentage)”.
- L475. Delete “also”
- L564. Consider use “internal variability” than “noise”.
- L547-548. While I agree it would be nice to have models use similar or save vertical grid/resolution, the adoption of vertical grids is a decision made by each modeling groups in consideration of all factors, including numerical recipes. This recommendation/suggest is just an overkill. I would suggest delete this sentence. As an alternative approach, you may mention that differences in vertical grids can be a major contributing source of stratospheric H2O lifetime.
Citation: https://doi.org/10.5194/egusphere-2024-2866-RC1 -
RC2: 'Comment on egusphere-2024-2866', Anonymous Referee #2, 05 Nov 2024
Van’t Hoff et al. explores the effects of an implementation of a supersonic aircraft fleet on atmospheric composition, comparing results from four models and contrasting them to a previous study using different models but a similar emission scenario. The concerns over environmental effects of supersonic aircraft sparked research decades ago and given that these aircraft are now back on the agenda, the study is timely. The study is comprehensive and mostly well-written, but I still have some comments and questions to be addressed before I can recommend publication.
I think the reflection on the broader importance of this study can be improved. Many of the changes from implementing a supersonic fleet are provided with two significant digits but are quite small. While I understand quantification of radiative effects may be beyond the scope of the study, but a brief discussion about climate implications and/or e.g. placing the changes from implementation of supersonic aircraft into the context of the current impact of the fleet, would be good. Including the potential role of the reduced Nox emissions at lower altitudes, which I couldn’t really see discussed much. Strangely, the authors only mention H2O as a climate driver, not ozone.
In several places, the authors conclude that there is significant improvement in the model agreement compared to the older study. However, looking at their figures, in particular the vertical profiles, I don’t think this is a statement that can be made without further quantification or a definition of what the authors consider improvement. Moreover, given that the comparison of inter-model differences is limited by the fact that there are notable differences in model setup and inconsistent parameterizations, there is even a stated issue with one of the model’s treatment of H2O, I’m not convinced this “improved agreement” is something you want to trust. Finally, given that different models are used in the current and the old study, there’s a significant limit to how far the comparison of agreement can be pushed.
The study speaks of substantial improvements in the modeling of ozone chemistry as the reason for the improved agreement, but studies still show significant differences from observations as well as between models in the baseline representation of the atmospheric composition, and there are model differences in the assessment of the current aviation fleet. More complex chemistry does not always equal improved performance. The authors should consider including a (clearer) discussion of the baseline model performance and whether biases play a role in the differences when assessing emission scenarios.
The methods section would benefit from some streamlining and clarifications – see detailed comments below. The results section reports is comprehensive but a bit sloppy at times, and it would help the reader with a bit more precise language, e.g. making sure to say what the increases and decreases are relative to (e.g. “There is little change in terms of the H2O perturbation” – does this refer to change relative to the baseline or change compared to the difference between A1 and baseline), consistent use of +/- or not in front of numbers, etc. I also wonder if the current level of precision of the reported numbers is needed, or even warranted here, given model spread or if it rather distracts from the core message of model differences. Moreover, I think there is some repetition between section 3.6 and preceding sections that could be avoided (by combining the discussion and comparison with other studies into one section), hence improving readability (of a section that is quite long).
Specific comments:
Line 60: ozone change also has a climate impact – a bit strange to only speak of water vapor as having a climate effect.
Line 66-68: this is the case also for aviation in general, not just for a supersonic fleet – may want to make that point.
Line 70: use response instead of impact or specific impact of what
Line 84-86: it’s a bit unclear from this and the preceding sentence how you evaluate an improvement – please could you specify or use comparable units for the two set of numbers reported
Line 105: should be clear about why 2050 atmosphere means – as far as I can tell, there is not change in meteorological/climatic conditions to align with a projected SSP370 climate. Also, the description of what is done varies between models – for instance MOZART uses RCP 4.5, do all the models change e.g. CH4 concentration to SSP370 levels, GEOS-Chem only mentions volcanic emissions (what do other models do?), etc.
Line 110: it’s not quite clear to me why – is it because the implementation of supersonic aircraft happens on a different baseline aviation sector, i.e. SSP370? Would be good to clarify.
Table 1: a bit unclear. The left-hand side says “all aircraft” and the right “supersonic aircraft” but looking at it, it doesn’t seem like the all aircraft fuel consumption in the different scenarios adds up to the A0 plus whatever the introduction of a supersonic fleet does – which I’d expect from the titles. Please clarify.
Line 215: what exactly does fixing in the troposphere mean?
Line 215-216: this seems like a quite important problem, making me question inclusion of this the model in the model-mean values reported at this stage. But maybe the authors have mode confidence than is reflected by this wording – if so, I would suggest modifying
Line 226: again, unclear what a 2050 atmosphere means
Line 248: missing a +-sign? Also, text should specify what the percent number is relative to.
Line 253: rather, the change in emission cause
Line 260: without any quantitative measure or validation, and given the caveats and differences in model experimental setup listed above, I question whether you can and should confidently say that reduced model spread is a “considerable improvement”. If Suggest rephrasing to “considerably lower model spread” or “closer model agreement” Importantly, I do not agree that the model agreement seems to be smaller when looking at the spatially resolved figures. So this needs to be said in a more nuanced way.
Line 265: is 3.4 the model average? Not clear from the text/table
Table 3: for ozone, is it the full column or stratospheric only?
Line 277: can the authors explain why this occurs?
Fig 2: caption says in response to supersonic emissions scenario (A0) – should this be A1?
Line 294-295: what does extended vertical domain mean? That the models have a higher upper model layer? This seems an important bit of information when looking at a perturbation to the atmosphere that is so altitude dependent and some more detail would be helpful.
Line 306: please elaborate on the term “tropical pipe”
Line 311: is discrepancy the best word for describing model differences? Requires a baseline to relate it to? Maybe large spread instead
Line 396: not sure I’m convinced that EMAC responses can be said to be similar here, moreover, when looking at the vertical profiles there seems to be neither similarity nor substantially lower model spread than in the Grewe et al study. This applies to other places as well, where the word similar is used for the model comparison.
Line 398-399: I did not readily find out how the different models treat heterogenous chemistry, some have polar stratospheric clouds as well – some more detail for all models in the methods section would be helpful. Do all four models have aerosol treatment included?
Line 401: should be ozone layer’s self-healing effect? Not sure what is means for ozone to be self-healing… Also, what are smog forming processes when you talk about lower-stratosphere? I think it’s better and more clear to describe in terms of chemistry. And relate it to the NOx emission changes displayed in Fig1.
Line 417-418: the cited study is very old and much older than the Grewe et al study. Are there more recent model documentation or evaluation papers that can be used to support this statement? Alternatively, what improvements have been made to these models compared to the ones used in Grewe et al.
Citation: https://doi.org/10.5194/egusphere-2024-2866-RC2 - AC1: 'Response to referee comments', Irene Dedoussi, 19 Dec 2024
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 202 | 36 | 87 | 325 | 3 | 4 |
- HTML: 202
- PDF: 36
- XML: 87
- Total: 325
- BibTeX: 3
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1